How to evaluate the prognosis and develop overall treatment strategies of metachronous bilateral breast cancer (MBBC) remains confused in clinical. Here, we investigated the correlation between clonal evolution and clinical characteristics of MBBC; we aim to establish a novel prognostic model in these patients.

The data from Surveillance, Epidemiology, and End Results (SEER) database and the First Hospital of Jilin University were analyzed for breast cancer–specific cumulative mortality (BCCM) by competing risk model. Meanwhile, whole-exome sequencing was applied for 10 lesions acquired at spatial–temporal distinct regions of five patients from our own hospital to reconstruct clonal evolutionary characteristics of MBBC. Then, dimensional-reduction (DR) cumulative incidence function (CIF) curves of MBBC features were established on different point in diagnostic interval time, to build a novel DR nomogram.

Significant heterogeneity in genome and clinical features of MBBC was widespread. The mutational diversity of contralateral BC (CBC) was significantly higher than that in primary BC (PBC), and the most effective prognostic MATH ratio was significantly correlated with interval time (R² = 0.85, p< 0.05). In SEER cohort study (n = 13,304), the interval time was not only significantly affected the BCCM by multivariate analysis (p< 0.000) but determined the weight of clinical features (T/N stage, grade and ER status) on PBC and CBC in prognostic evaluation. Thus, clinical parameters after DR based on interval time were incorporated into the nomogram for prognostic predicting BCCM. Concordance index was 0.773 (95% CI, 0.769–0.776) in training cohort (n = 8,869), and 0.819 (95% CI, 0.813–0.826) in validation cohort (n = 4,435).

Bilateral heterogeneous characteristics and interval time were determinant prognostic factors of MBBC. The DR prognostic nomogram may help clinicians in prognostic evaluation and decision making.

如何评估异时性双侧乳腺癌（MBBC）的预后并制定整体治疗策略在临床上仍存在困惑。在这里，我们研究了 MBBC 的克隆进化与临床特征之间的相关性；我们的目标是在这些患者中建立一种新的预后模型。

采用竞争风险模型对来自监测、流行病学和最终结果（SEER）数据库和吉林大学第一医院的数据进行乳腺癌特异性累积死亡率（BCCM）分析。同时，对我院5名患者时空不同区域采集的10个病灶进行全外显子组测序，重建MBBC的克隆进化特征。然后，在诊断间隔时间的不同点建立MBBC特征的降维（DR）累积发生函数（CIF）曲线，构建新的DR列线图。

MBBC 基因组和临床特征的显着异质性很普遍。对侧BC（CBC）的突变多样性显着高于原发性BC（PBC），最有效的预后MATH比与间隔时间显着相关。R² = 0.85，p<0.05)。在 SEER 队列研究中（n= 13,304），通过多变量分析，间隔时间不仅显着影响 BCCM（p< 0.000），但在预后评估中确定了 PBC 和 CBC 临床特征（T/N 分期、等级和 ER 状态）的权重。因此，基于间隔时间的 DR 后的临床参数被纳入列线图中，用于预测 BCCM 的预后。训练队列的一致性指数为 0.773（95% CI，0.769-0.776）（n= 8,869) 和 0.819 (95% CI, 0.813–0.826) 在验证队列 (n= 4,435)。

双侧异质性特征和间隔时间是 MBBC 的决定性预后因素。DR 预后列线图可以帮助临床医生进行预后评估和决策。